feat(transformer): numeric separator plugin.

[rzvxa]

es2021 numeric separator

[codspeed-hq]

CodSpeed Performance Report

Merging #2795 will improve performances by 3.13%

_{Comparing rzvxa:num-sep (dcbde1d) with main (56493bd)}

Summary

⚡ 1 improvements
✅ 33 untouched benchmarks

Benchmarks breakdown

	Benchmark	main	rzvxa:num-sep	Change
⚡	parser_napi[checker.ts]	240.7 ms	233.4 ms	+3.13%

[Dunqing]

You need to update the snapshot. Just run just coverage

[rzvxa]

You need to update the snapshot. Just run just coverage

Thanks, Yeah I'm aware of the snapshot being different. My issue with this one is that I'm only passing one of 2 tests and since it is such a small thing I can't put my finger on the thing that causes this issue, Have to do some tracing to figure it out.

  # All Passed:
-* babel-plugin-transform-numeric-separator
 * babel-plugin-transform-optional-catch-binding
 * babel-plugin-transform-json-strings
 * babel-plugin-transform-shorthand-properties
@@ -515,6 +514,9 @@ Passed: 352/1369
 # babel-plugin-transform-logical-assignment-operators (5/6)
 * logical-assignment/null-coalescing/input.js
 
+# babel-plugin-transform-numeric-separator (1/2)
+* used-with-transform-es2015-literals/input.js
+
 # babel-plugin-transform-export-namespace-from (0/4)
 * export-namespace/namespace-default/input.mjs
 * export-namespace/namespace-es6/input.mjs

It seems like the snapshot already has these tests passing, Since they were implemented in codegen, It is broken since I'm passing 1 of 2 instead of 2 of 2 so updating snapshots shouldn't be necessary when I get it working.

[rzvxa]

@Dunqing Sorry to bother you, Do you have any idea about this? I'm generating the same output as Babel but the test fails. Is there anything that I'm missing?

I'm getting this:

Original:

1_000;
0xA0_B0_C0;
0b1010_0001_1000_0101;
0o0_6_6_6;


Transformed:

1000;
0xA0B0C0;
41349;
438;

and Babel produces this:

1000;
0xA0B0C0;
41349;
438;

Here is the config used for the babel:

{
  "plugins": [
    "@babel/plugin-transform-numeric-separator",
    "@babel/plugin-transform-literals"
  ]
}

[Dunqing]

@Dunqing Sorry to bother you, Do you have any idea about this? I'm generating the same output as Babel but the test fails. Is there anything that I'm missing?

I'm getting this:

Original:

1_000;
0xA0_B0_C0;
0b1010_0001_1000_0101;
0o0_6_6_6;


Transformed:

1000;
0xA0B0C0;
41349;
438;

and Babel produces this:

1000;
0xA0B0C0;
41349;
438;

Here is the config used for the babel:

{
  "plugins": [
    "@babel/plugin-transform-numeric-separator",
    "@babel/plugin-transform-literals"
  ]
}

I will take a look soon

[Dunqing]

There are two problems that need to be solved

1. The test lacks an output file, which is a babel problem fix: incorrect numeric-seperator tests babel/babel#16381
2. Due to the lack of an output file, oxc transform_confromance will use the input file as the output file fix(tasks/transform-conformance) when the output file does not exist, the output content should be empty #2808

[rzvxa]

@Dunqing Wow great work! It literally took me a day or so to look at the same code wondering why is it wrong, Never would've thought that the test case itself was incomplete.

I guess with this knowledge, We can think of this and #2797 as passing the tests since they are producing the same output as Babel.

[Dunqing]

I think we can merge this after you have cleaned up the code and fixed the conflict.

[Dunqing]

Thank you!!